100Hz Hum in Lehmann Linear clone

[PsychoM]

Anything below 80 degrees Celsius is ok. Grab your thermometer because your fingers can lie - metal parts tend to feel hot.

The heatsinks are all at about 65 degrees Celsius. I also measured bias voltage and it's the same as on the original Lehmann.
So they're not running too hot (still, I wouldn't want to have a serious power amplifier running that hot, unless I'd want to use it as a heater too).

Probably, the heatsinks aren't of great quality, after all it was a very cheap kit. That might explain the temperature too.

thanks again

 

[jitter]

The somewhat high temperature is the result of the size of the heatsinks combined with the amount of power the output transistors are dissipating. It looks like the size of those heatsinks is only adequate.

If I were to put this board in a small casing or one without ventilation slots, I'd probably mount the transistors upright on a single bigger heatsink, with insulation pads, of course. Mounting the transistors upright clears enough space for such a heatsink.
I did something similar with my diy class A headphone amplifier with eight BD139/B140s. The heatsink is overdimensioned for the purpose, but this way I get away with using a compact case without any ventilation slots.

Edit: and a tip: use thermal paste very sparingly. There should only be enough on the backside of the transistors to fill any roughness to improve thermal contact. But this thermal grease has a thermal resistance of its own, so using too much of it (like in your pictures) might actually be counter productive.

 

[flyingfishtw]

I've never try the clone one myself, it just looks so interesting that designer put so much effort to reduce ripples of power supply which is good for the diamond buffer, but ignored more important part, ground voltage difference,which might get into signal 100%.

few years ago i've done a similar configuration with diamond buffer, i made the same fault as lehmann amp. The output quiescent current will gradually increasing, eventually it might collapse if the room temperature or quiescent is set too high. that can be easily fixed by couple transistor together. which will make the board even more ugly, but at least it will be super stable.

So lesson learned, now when I layout a board, I consider what the components requirements more than the looking.

 

[PsychoM]

Edit: and a tip: use thermal paste very sparingly. There should only be enough on the backside of the transistors to fill any roughness to improve thermal contact. But this thermal grease has a thermal resistance of its own, so using too much of it (like in your pictures) might actually be counter productive.

I know there's too much thermal paste on 2 of the transistors, but that's because I had to reinstall the heatsinks several times. Those two heatsinks were too long, they shorted the pins of the transistors and I had to grind a few mm off...

I tested the amp for a long time, and the temperature seems to be steady at 65 degrees Celsius. But that's at room temperature (+-22 degrees) and in open air.

Maybe I'll have to use a fan if I use a case. Even a case with large heat vents might not be enough.

 

[wwenze]

It's unfortunate that the designer designed the transistors to be lying down; if they were standing you could choose better heatsinks.

 

[PsychoM]

Final update on this one. Amp is up and running. Still getting hot, but it never seems to go over 65 degrees, even in the case I designed to accompany its little brother, a TPA3116 amp.
Unfortunately, I accidentally reversed the ventilation holes on the sides, but it seems the temperature is still OK. I need to get some new metal output jacks and an extra pair of RCA jacks, and it's finished.



Thanks again for the help

 

[Clean6eR]

hi,

i have one of the LCL clones and it has a constant hum, i tried the cutting the ground and running a ground as shown in the pictures on page 2 here but it has made no difference at all, its not any worse though so not all bad, any ideas what might be the cause and solution?

 

[manso]

If your regulator is same as shown on in post 1, use a bigger cap to bypass regulator adjust pin, .15uf wima cap is too little, use min 10 uf and better 22uf. Im surprised no one noticed this, those regs are probably oscilating with such big electrolytics on their output.

 

[murrays]

Transformer too close to signal wires or circuit?

 

[Clean6eR]

If your regulator is same as shown on in post 1, use a bigger cap to bypass regulator adjust pin, .15uf wima cap is too little, use min 10 uf and better 22uf. Im surprised no one noticed this, those regs are probably oscilating with such big electrolytics on their output.

my amp has "HUM_AUDIO PHONE BUFFER VER11" written on it and it is simular in layout to the pictures on page 2 but the components have different labels.

my LM317T is U3 and the LM337T is U2!

i presume you mean i need to change C210 and C209 to 22uf ones. on the picture on the previous page "PsychoM" has six fat red caps in that position, mine a red WIMA ones that i can't read the label to clearly.

do i just switch the middle 2 out with the better cap then im all happy?

can you link to any good caps that you think might be small enough and still meet the voltage needed for them? what voltage is on them? (my multi meter is in a shed outside and I'm nice and warm inside crying atop of the amp

I'm not great with electronics but im trying so thank you for your suggestions and help.

 

[Clean6eR]

Transformer too close to signal wires or circuit?

Well its in s case that's the same dimensions as the BCL, the transformer is close to the input lines, but i have routed them as far away as they can fit.

i thought it might be the transformer but i also don't have much experience with audio engineering so my thought means nothing!

its a 30VA toroidal one (no brand printed on it) like found in the BCL but i was wondering if i slip a R core in place of it would it reduce noise or perform better for this amp?

has anyone worked out the finest transformer for this clone?

 

[mrmisi]

I see this topic is aged, but maybe someone can get helped to stop annoying 100 Hz buzz comes from Lehmann headphone amp clones.
Since I prefer A class, and amplifiers are put together by myself, a logic decision was LA clone to order from ebay.
It was hard to imagine that any basic groundig issues remained in a clone board from LA, but yes.
The fisrt trial was disappointing. A permanent 100 Hz noise was sit on the output, independently from how the volume control was set. Since it was not a tipical 50Hz noise, then became obvious, nothing to do on transformer or cable positioning. I called my best friend, Fluke 105B scopemeter and started to trace back the source of groundloop.
The output of opamap was clean, no 100Hz noise, dead silent, offset voltage was hard to measure. Ok, then lets see the transistor output. The 100Hz gremlins should be hidden someweher here. After all important Dc and Ac routine measure of output stage I was a bit stuck. By checking again the schematic and Pcb layout, and noiseless opamap section, I recognized that opamap and output stage groundig points are not routed back directly to the same common point. And yes, by cuting the headphone jack socket groundig point on Pcb directly under the socket ( you wont see It after placing socket back, no visual footprint) and retourning back the soldering point from here with a piece of cable on the bottom side to grounding point of decoupling capacitors of opamap, you will have a dead silent, groundloop free amp.

 

[withmatt]

I see this topic is aged, but maybe someone can get helped to stop annoying 100 Hz buzz comes from Lehmann headphone amp clones.
Since I prefer A class, and amplifiers are put together by myself, a logic decision was LA clone to order from ebay.
It was hard to imagine that any basic groundig issues remained in a clone board from LA, but yes.
The fisrt trial was disappointing. A permanent 100 Hz noise was sit on the output, independently from how the volume control was set. Since it was not a tipical 50Hz noise, then became obvious, nothing to do on transformer or cable positioning. I called my best friend, Fluke 105B scopemeter and started to trace back the source of groundloop.
The output of opamap was clean, no 100Hz noise, dead silent, offset voltage was hard to measure. Ok, then lets see the transistor output. The 100Hz gremlins should be hidden someweher here. After all important Dc and Ac routine measure of output stage I was a bit stuck. By checking again the schematic and Pcb layout, and noiseless opamap section, I recognized that opamap and output stage groundig points are not routed back directly to the same common point. And yes, by cuting the headphone jack socket groundig point on Pcb directly under the socket ( you wont see It after placing socket back, no visual footprint) and retourning back the soldering point from here with a piece of cable on the bottom side to grounding point of decoupling capacitors of opamap, you will have a dead silent, groundloop free amp.

I had very similar problem. Hum with my Lehmann clone was not so big, bu noticeable with low impedance headphones - with my favourite KSC35.
It happened even without opamp. It happened with mod introduced in this thread.. so I started investigation of crappy GND design.

In my case, hum was gone when I connected GND point from potentiometer GND pin to headphone jack GND point - where GND is separated on top layer. It seems my solution is very similar to yours.

Now it is dead-silence and lovely!

 

[mrmisi]

I guess many alternative solutions exist and depend on which PCB version you have. The one is purchased just 2 weeks ago as improved version requires more modification, than randomly touching the grounding points and hoping the best which works in some cases. I guess you were more lucky with yours. The latest version's output jack is grounded to "high current" loop which not so optimal way. Making small signal audio ground together with high current one looks even worst, and might degrade already working parts of the layout. The more optimal solution to place back all players in this layout into proper place, and "let them to kick the ball into nett located on proper side." I am not a fan of surgeon of the PCB but sometimes need to cut if brings ultimate solutions and no visible remarks remain on.

 

[DarkskyZ]

Sorry to dig the old thread out but I do too have humming noises on mine (Lovely Cube from alicemagicbox back from 2010). I left it in a box for years but I just bought a new pair of nice cans and I decided to try it again. I realised why I left it alone, humming all day long is really a turn off...

Anyway, I already did part of flyingfishtw's recommended mods, I wired everything as described but I still have to cut the ground plane as the hum is still there. I just don't really know how to do it. I am not really into electronics and haven't even soldered anything for years before these few wires on the amp.

Do I really need to scar the board all along the way like PsychoM did ?
What kind of tool would be suitable ? exact knife maybe ?

Any guidance would be hugely appreciated. I already have ordered a Schiit Magni 3 to replace this one anyway but I would love to put the Linear back in working order anyway.

Thanks !

Edit : sorry for the possible double post, this is my first post here and I missed the fact that it had to be approved by a mod before being made public...

 

[sgrossklass]

What kind of tool would be suitable ? exact knife maybe ?

That should do.

If there is a ground loop on the board, you don't get around some cutting action I'm afraid.

Should you ever come across a PC with an interference-ridden front headphone output but perfectly clean rear output, the solution would be along the same lines (cutting the connection between USB ground and audio ground on the front panel board).

 

[DarkskyZ]

Thanks for answering sgrossklass, I see what the goal of this is but what I don't really get is how it's achieved in the particular case of the Lehmann Linear clone by scarring the board

Here is mine running (and humming...)

How do I get to scrape the board like PsychoM did his with all the components on ? Anybody care to share how they did theirs ?

I really don't get it. I do feel really dumb I have to admit.

 

[sgrossklass]

You would probably have to unsolder the rightmost 470µ/25V cap to gain access to the area. (If the regulator heatsink is in the way, I'd try unscrewing and bending it up temporarily.) It doesn't have to be a super-wide cut, just as long as the two pieces are no longer connected. It would probably have been smart to do the cutting first, otherwise you'd have to temporarily disconnect the mod wires again to check success with a multimeter. (Well, I guess you could still use your ears...)

Once the board is successfully modified, it would be nice to seal the cut back up... maybe some varnish or something. There's probably a million options, I guess clear fingernail polish would be one of them...

 

[DarkskyZ]

Thanks again for your kind help

I actually bought the Lovely Cube fully assembled so I didn't have a chance to modify anything beforehand.

With your advice, I just braced myself and did the thing. It seems I managed to scrape it enough without unsoldering anything as it is now silent when it should be ! No more hum it seems.
I'll still let it run on full volume for a few hours as I noticed the hum became more and more present with the amp warming up.

The board now looks awful but who cares anyway.

I'll put some kind of varnish on the "scar" tomorrow.

Thanks again.

 

[WoodyLuvr]

Thanks again for your kind help

I actually bought the Lovely Cube fully assembled so I didn't have a chance to modify anything beforehand.

With your advice, I just braced myself and did the thing. It seems I managed to scrape it enough without unsoldering anything as it is now silent when it should be ! No more hum it seems.
I'll still let it run on full volume for a few hours as I noticed the hum became more and more present with the amp warming up.

The board now looks awful but who cares anyway.

I'll put some kind of varnish on the "scar" tomorrow.

Thanks again.

Outstanding repair there! Please snap and send a pic if you can.